Introduction
Although superficial nonampullary duodenal epithelial tumors (SNADETs) were previously
considered rare, their recognition has been increasing following an improvement in
endoscopic accuracy [1 ]
[2 ]
[3 ]
[4 ]
[5 ]. The opportunities to perform endoscopic resection of SNADETs have therefore recently
increased; however, it is considered extremely difficult because of the technical
challenges posed by the complex anatomic features of the duodenum [2 ]
[4 ]
[5 ]. Recently, the safety and efficacy of underwater endoscopic mucosal resection (UEMR)
have been increasingly reported [6 ]
[7 ]
[8 ]
[9 ]
[10 ]
[11 ]
[12 ]
[13 ]
[14 ]
[15 ]. UEMR is a water immersion method leading to easy snaring and a low perforation
risk [2 ]
[7 ]
[11 ]
[12 ]
[13 ]
[14 ]. Therefore, UEMR is a favorable treatment and results in high resection and low
adverse event (AE) rates [4 ]
[7 ]
[8 ]
[9 ]
[13 ]
[14 ]
[15 ]. However, continuously maintaining water in the lumen is difficult, leading to large-volume
water injection and a long procedure time [6 ]
[13 ].
Recently, a novel gelatinous liquid (Viscoclear; Otsuka Pharmaceuticals Factory, Inc.,
Tokushima, Japan) was developed to secure the visual field during endoscopy [16 ]
[17 ]. This gelatinous liquid comprises xanthan gum, locust bean gum, and glycerin [16 ]. The viscoelastic properties of the gelatinous liquid enable continuous filling
of the lumen and prevent its mixing with fresh blood or feces, resulting in a favorable
space for endoscopic visualization and treatment [16 ]
[17 ]. Yano et al. reported that use of an immersing gelatinous liquid during endoscopy
resulted in a favorable visual field and named this method gel immersion endoscopy
[16 ]
[17 ]
[18 ]. Applying these features, we devised gel immersion endoscopic resection (GIER),
a gel-based EMR technique, as an alternative method to water-based UEMR, and reported
its efficacy [19 ].
Although GIER may effectively treat SNADETs, there have been no reports comparing
GIER and UEMR. Therefore, we retrospectively investigated the short-term clinical
outcomes of GIER as a treatment for SNADETs as compared with UEMR.
Methods
Study design and participants
We retrospectively reviewed the outcomes of 40 consecutive SNADET cases treated by
GIER and UEMR in 35 patients between April 2019 and December 2021 at Asahi General
Hospital, Japan. Because the gelatinous liquid was launched in the Japanese market
in October 2020, the 18 consecutive cases before the launch (between April 2019 and
September 2020) were treated with UEMR, whereas the 22 consecutive cases after the
launch (between October 2020 and December 2021) were treated with GIER.
The institutional review board of Asahi General Hospital approved this retrospective
study (16 March 2021). Written informed consent for the procedure was obtained from
all the patients. Endoscopic resection was indicated for SNADETs ≤ 20 mm in size.
Procedures
Two board-certified fellows of the Japan Gastroenterological Endoscopy Society and
one trainee performed all of the procedures. The patients were admitted 1 day prior
to endoscopic resection or on the day of the procedure, and remained in the hospital
for approximately 5 days. A gastroscope (GIF-Q260J; Olympus Medical Systems, Tokyo,
Japan) was used with carbon dioxide insufflation.
For GIER, the gelatinous liquid was injected gently using a syringe, continuously
filling the lumen; for UEMR, water was injected. Carbon dioxide insufflation was discontinued
during immersion and resection. The lesion was captured using a 10-mm (AG-5076–241023;
Hangzhou AGS MedTech Co., Ltd., Zhejiang, China), 13-mm (M00562451; Boston Scientific,
Massachusetts, United States), 15-mm (AG-5076–241523; Hangzhou AGS MedTech Co., Ltd.),
and 20-mm electrocautery snare (M00561831; Boston Scientific) as appropriate under
liquid immersion. The current was passed using a high frequency generator (VIO300D;
Erbe Elektromedizin, Tübingen, Germany). The high frequency generator was used with
dry-cut mode (effect 4, 40 W) for mucosal resection. Prophylactic clip closure was
performed for all cases using endoclips (HX-610–135S; Olympus).
A second-look endoscopy was routinely performed in all cases. All patients received
20 mg intravenous omeprazole for the first 2 days, beginning on the day of the procedure.
Thereafter, 20 mg rabeprazole daily was prescribed for approximately 30 days. Treatment
of a patient using GIER is shown in [Fig. 1 ] and [Video 1 ].
Fig. 1 Example of gel immersion endoscopic resection of a superficial nonampullary duodenal
epithelial tumor. a–d Endoscopic views showing: a,b a superficial 11-mm elevated and depressed lesion in the second part of the duodenum;
c the favorable visual field after the lesion had been immersed in a gelatinous liquid;
d a sufficient horizontal margin being easily confirmed with the lesion under the gelatinous
liquid. e,f Macroscopic appearance of the en bloc resected lesion after it had been easily captured
using an electrocautery snare, resulting in a specimen of 25 × 10 mm, containing a
tumor of 11 × 8 mm that was pathologically diagnosed as a well-differentiated adenocarcinoma
with negative margins.
Video 1 Demonstration of the gel immersion endoscopic resection technique for a superficial
nonampullary duodenal epithelial tumor.
The resected specimens were embedded in paraffin and subjected to staining with hematoxylin
and eosin for histopathological diagnosis. R0 resection was defined as an en bloc
resection where both the horizontal and vertical margins were negative. For en bloc
resection specimens, we considered adenomas with unclear lateral margins and expected
lateral burning effect as negative margins.
Outcome definitions
The primary outcome was procedure time, defined as the time from liquid immersion
into the lumen to resection. The secondary outcomes were en bloc resection rate, R0
resection rate, tumor size, specimen size, and AEs, such as delayed bleeding and perforation.
Delayed bleeding was defined as hematemesis or melena requiring endoscopic hemostasis
or blood transfusion, or leading to a reduction ≥ 2 g/dL in hemoglobin level.
Statistical analyses
Continuous variables were reported as the median and range, and compared using the
Mann–Whitney–Wilcoxon test. Differences in the categorical variables were analyzed
using the likelihood ratio test and Fisher’s exact test. These analyses were conducted
at the Clinical Research Support Center of Asahi General Hospital. A two-sided P value of < 0.05 was considered statistically significant. All statistical analyses
were performed using the JMP version 16.1.0 software (SAS Institute Japan, Tokyo,
Japan).
Results
Patient and lesion characteristics in the GIER and UEMR groups
Overall, 35 patients were treated with 40 consecutive endoscopic resection procedures.
UEMR was performed for the first 18 cases, while GIER was performed for the latter
22 cases. The comparison of patient and lesion characteristics between GIER and UEMR
groups is presented in [Table 1 ]. No significant differences were observed in the patient and lesion characteristics.
Table 1
Patient and lesion characteristics in the gel immersion endoscopic resection (GIER)
and underwater endoscopic mucosal resection (UEMR) groups.
GIER
UEMR
P value
Patients, n
19
16
Age, median (range), years
67 (45–81)
63 (35–80)
0.25[1 ]
Sex, n (%)
0.72[2 ]
13 (68.4)
12 (75.0)
6 (31.6)
4 (25.0)
Lesions, n
22
18
Location, n (%)
0.43[3 ]
3 (13.6)
1 (5.6)
16 (72.7)
16 (88.9)
3 (13.6)
1 (5.6)
Macroscopic type, n (%)
0.19[3 ]
6 (27.3)
1 (5.6)
9 (40.9)
10 (55.6)
3 (13.6)
5 (27.8)
4 (18.2)
2 (11.1)
0-I, protruding; 0-IIa, superficial elevated; 0-IIc, superficial depressed.
1 Mann–Whitney–Wilcoxon test.
2 Fisher’s exact test.
3 Likelihood ratio test.
Treatment outcomes in the GIER and UEMR groups
The median (range) procedure time was significantly shorter in the GIER group than
in the UEMR group (2.75 [1–3.5] minutes vs. 3 [2 ]
[3 ]
[4 ]
[5 ]
[6 ]
[7 ]
[8 ]
[9 ]
[10 ] minutes; P = 0.01) ([Table 2 ]). The en bloc resection rate was 100 % (95 %CI 85.1 %–100 %) in the GIER group,
but only 83.3 % (95 %CI 60.8 %–94.2 %) in the UEMR group (P = 0.08). The R0 resection rate was significantly higher in the GIER group than in
the UEMR group (95.5 % [95 %CI 78.2 %–99.2 %] vs. 66.7 % [95 %CI 43.7 %–83.7 %]; P = 0.03). The median (range) specimen size was larger in the GIER group than in the
UEMR group (14 [7–27] mm vs. 7.5 [4–11] mm; P < 0.001). Tumor size and histological type were not significantly different between
the groups. No AEs, such as perforation and delayed bleeding, were observed in the
groups. The median (range) amount of gelatinous liquid used during a procedure was
180 (60–300) mL in the GIER group.
Table 2
Treatment outcomes of the gel immersion endoscopic resection (GIER) and underwater
endoscopic mucosal resection (UEMR) groups.
GIER
UEMR
P value
Tumor size, median (range), mm
7 (2–18)
5 (3–10)
0.48[1 ]
Specimen size, median (range), mm
14 (7–27)
7.5 (4–11)
< 0.001[1 ]
Pathology, n (%)
0.18[3 ]
1 (4.6)
4 (22.2)
18 (81.8)
13 (72.2)
3 (13.6)
1 (5.6)
Procedure time, median (range), minutes
2.75 (1–3.5)
3 (2–10)
0.01[1 ]
En bloc resection, n (% [95 %CI])
22 (100 [85.1 %–100 %])
15 (83.3 [60.8 %–94.2 %])
0.08[2 ]
R0 resection, n (% [95 %CI])
21 (95.5 [78.2 %–99.2 %])
12 (66.7 [43.7 %–83.7 %])
0.03[2 ]
Adverse events, n
0 (0)
0 (0)
Liquid consumption, median (range), mL
180 (60–300)
No data
Discussion
This study observed that GIER, a form of gel immersion endoscopy, resulted in a short
procedure time. Moreover, GIER enabled the resection of a larger specimen size and
led to an increased R0 resection rate.
We observed that the procedure time was significantly shorter in the GIER group than
in the UEMR group. A Japanese multicenter prospective study reported that the procedure
time for treatment with UEMR was 5.4 (SD 4.3) minutes [14 ]. Moreover, the previous reports comparing UEMR and conventional EMR revealed that
the procedure times were 4–7 minutes and 9.5–12 minutes, respectively [12 ]
[13 ]
[20 ]. We believe that our procedure time in patients treated with GIER is markedly shorter
than previously reported for treatment with UEMR and conventional EMR. The gelatinous
liquid used in GIER allows for easier maintenance of continuous lumen filling compared
with water, and an additional infusion is unnecessary, leading to a short infusion
time [19 ]. Furthermore, the viscoelastic properties of the gelatinous liquid enable a favorable
transparent visual field to be maintained, resulting in easy snaring [19 ]. These factors might have led to a short procedure time.
The en bloc resection rate tended to be higher in the GIER group, while the R0 resection
rate was significantly higher in the GIER group than in the UEMR group. Moreover,
the specimen size was larger in the GIER group than in the UEMR group. The previous
reports comparing UEMR and conventional EMR revealed that the en bloc resection rates
were 75.4 %–96.4 % and 60.9 %–96.4 %, respectively [11 ]
[12 ]
[13 ]
[15 ]
[20 ]. Moreover, R0 resection rates were 50.8 %–76.1 % and 34.8 %–79.6 %, respectively
[11 ]
[12 ]
[13 ]
[15 ]
[20 ]. In comparison with these reports, we believe that the en bloc and R0 resection
rates for treatment with GIER are favorable. As mentioned previously, the favorable
visual field under the gelatinous liquid might assist with snaring by allowing a sufficient
horizontal margin to be determined, leading to a high resection rate and a large specimen.
A Japanese multicenter retrospective study reported that the en bloc resection rate
was significantly lower in patients undergoing UEMR compared with conventional EMR,
especially for large lesions [20 ]. Furthermore, a single-center study revealed that UEMR had lower en bloc and R0
resection rates compared with conventional EMR, although the rate of conversion to
endoscopic submucosal dissection was reduced [11 ]. We believe that GIER leads to a large specimen with en bloc resection and can eliminate
the drawbacks of UEMR. Moreover, as these advantages may be applicable to the colon,
validation through further research is warranted.
When performing GIER, it is necessary to consider the electrical conduction properties
of the gelatinous liquid, as well as the risks of current transmission to the wall
and delayed transmural thermal injury. As the gelatinous liquid does not contain electrolytes,
it has the advantage of not discharging when monopolar devices are used [16 ]
[17 ]. Yano et al. established appropriate electrical conductivity using rat liver and
developed the dedicated gel [16 ]. They also verified the efficacy of electrocoagulation under the gelatinous liquid
with monopolar hemostatic forceps in a porcine model and revealed that severe coagulative
necrosis was not observed in the muscularis but was limited to the mucosa [16 ]. Therefore, we believe that the efficacy of electrocoagulation is favorable and
that the risks of current transmission to the wall and delayed transmural thermal
injury are low.
Although we propose GIER as a novel replacement method for UEMR, GIER has some drawbacks.
First, the gelatinous liquid costs 2000 Japanese Yen per 200 mL, which is more expensive
than water. Therefore, a validation of GIER cost-effectiveness is warranted. Second,
the specific gravity of the gelatinous liquid is heavier than that of water, which
may cause loss of buoyancy of the mucosal and submucosal layers, thereby reducing
the advantage offered by the water immersion method. Furthermore, the viscosity of
the gelatinous liquid may possibly affect the opening and closing of the snare. These
issues might make snaring difficult and lead to perforation and a low resection rate;
however, our previous reports revealed that the buoyancy of the mucosal and submucosal
layers was favorable and similar to that in underwater endoscopy based on endoscopic
ultrasonography findings [19 ]. Additionally, there was no difficulty in manipulating the snare. In fact, the resection
rate was high in the present study and no cases of perforation were observed.
This study has several limitations. First, this was a single-center retrospective
pilot study with a small sample size. Furthermore, some patients were included several
times. As a result, we cannot exclude the possibility of statistical dependencies.
It should be noted however that GIER is a novel treatment method, and there are no
case–control studies or comparative studies with regard to the treatment of SNADET
with GIER. Second, the impact of a learning curve and device improvement cannot be
denied. In this study, UEMR was performed for the first 18 cases, while GIER was performed
for the latter 22 cases. The improvement in technique owing to the learning curve
and the use of a novel snare might have led to favorable outcomes in the GIER group. Third,
the consumption of water was not recorded in the UEMR group. Therefore, it is unclear
whether GIER really reduced the liquid consumption compared with UEMR. Finally, long-term
clinical outcomes, such as local residual recurrence, were not validated. The evaluation
of whether GIER is definitely an effective treatment requires the verification of
long-term outcomes, and additional time is needed.
In conclusion, the gelatinous liquid enables continuous lumen filling and secures
a favorable visual field. GIER is a method that applies these features and might result
in a short and safe procedure. Therefore, GIER has the potential to be a novel treatment
method for SNADETs. To confirm its efficacy, multicenter prospective studies on the
short- and long-term clinical outcomes are needed.
